157
J. Sep. Sci. 2015, 38, 157–163
Yong Liu Dong Zhang Shiwu Pang Yanyun Liu Yu Shang Key Laboratory of Advanced Civil Engineering Materials, School of Materials Science and Engineering, Tongji University, Shanghai, P. R. China Received September 11, 2014 Revised October 17, 2014 Accepted October 20, 2014
Research Article
Size separation of graphene oxide using preparative free-flow electrophoresis Graphene oxide nanosheets often bear a wide size distribution. However, it is critical to have nanosheets with narrow size distribution for their unique size-dependent physiochemical properties, and nanosheets with a narrow size distribution are the cornerstones for application. Therefore, efficient separation methods of graphene nanosheets have been given considerable attention in many scientific areas recently. Free-flow electrophoresis is extensively used in the separation and purification of biological molecules with continuous flow separation. The charged graphene oxide nanosheets to some extent are very close in size to biological molecules and share similarity in motion behavior in an electric field. Thus, in the present work, we present a new and simple means to separate graphene oxide nanosheets into more mono-dispersed size groups by using the free-flow electrophoresis technique. By optimizing the separation conditions, we were able to obtain graphene oxide sheets with narrow size distribution. The separated samples were characterized by atomic force microscopy, and the size measurements were made by using the software “Image Pro Plus.” In addition, a brief discussion is also given into the theoretic background of the separation of graphene oxide according to the size by the technique of preparative free-flow electrophoresis.
Keywords: Free-flow electrophoresis / Graphene oxide DOI 10.1002/jssc.201401000
Additional supporting information may be found in the online version of this article at the publisher’s web-site
1 Introduction Graphene, a single layer of carbon atoms densely packed into a benzene-ring structure [1], has received considerable attention due to its unique electronic properties, high transparency, flexible structure, large specific surface area, and good thermal conduction [2, 3] since Geim et al. [1] made a breakthrough in the preparation of graphene by mechanical exfoliation. The peculiar chemical and physical properties of graphene are of great interest for potential applications in various fields, such as composite materials, batteries, sensors, solar cells and photo catalysts, and many others [4–6]. It is well known that the shape and size of graphene nanostructure dictate the electrical, optical, magnetic, and chemical properties because of the edge states and quantum Correspondence: Professor Dong Zhang, Key Laboratory of Advanced Civil Engineering Materials, School of Materials Science and Engineering, Tongji University, 4800 CaoAn Road, Shanghai 201804, P. R. China E-mail: [email protected] Fax: +86-21-69582144
Abbreviations: AFM, atomic force microscopy; FFE, free-flow electrophoresis; GO, graphene oxide
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
confinement [7–10]. Therefore, the key to controlling the properties mostly lies in the control of the size and shape of graphene. Unfortunately, current preparative processes of graphene have generally not generated a uniform size of nanosheets. In addition, graphene is poorly soluble in water or polar organic solvents, which makes it more difficult to further process to control the size and shape of the graphene nanosheets. Unlike hydrophobic graphene, graphene oxide (GO) is hydrophilic due to the oxygencontaining functional groups on the sheet surface, which gives the GO sheets good solubility in solvents and provides fertile opportunities for the processability of this nanomaterial [11]. Furthermore, GO carries sufficient functional groups (e.g. epoxides, hydroxyls, carboxylic acids) on the sheet surface. Ionization of the functional groups will occur in solution, and thus the GO sheets are negatively charged. In addition, GO has attracted intense interest for its use as a precursor material for the mass production of graphene-based materials, which hold great potential in various applications. Considering their unique size and/or shape-dependent properties, it is highly important to obtain GO nanosheets with a narrow size and shape dispersion. Many attempts have been made toward solving the above-mentioned issues such as magnetic stirring and
www.jss-journal.com
158
J. Sep. Sci. 2015, 38, 157–163
Y. Liu et al.
centrifugation [12], isopycnic density gradient ultracentrifugation [13, 14], and the density gradient ultracentrifugation rate [15]. However, it is not favorable to use centrifugation for separating GO on a large scale only if super-magnum centrifuge is given. It would be laborious and time-consuming if large amounts of separated samples are needed as it is done on a batch-by-batch basis, and lacks continuous separation capacity. Electrophoresis has been extensively used in the separation and purification of biological molecules with remarkable resolution [16]. Surprisingly, this technology has gained little exposure until recently in the separation of nanomaterials [17,18]. CE, a liquid-phase separation technique, has been reported for the separation of GO [19, 20]. In this technique, capillary as separation channel (micron scale), and high direct current (DC) electric voltage (about 30 kV) as the main driving force together achieve the analysis; CE lacks continuous separation capacity as separation is done on a batch-by-batch basis [21]. The CE method is done on a miniscule scale. Therefore, finding a fast, effective, and continuous method of GO separation remains a practical problem of critical importance. Free-flow electrophoresis (FFE), another little known electrophoresis technique, has evolved into one of the most promising techniques for preparative-scale separation in the field of biochemical technology [21–28]. FFE is a pure-liquid electrophoresis for both analytical and preparative separation in biochemical technology with the advantages of controllability, gentle conditions, high recovery, and complete preservation of activity. In FFE, pressure is used to drive a sample stream through a planar separation channel. An electric field is applied perpendicularly to the direction of carrier buffer flow while a sample solution is injected continuously into the carrier buffer as a narrow band to deflect sample into distinct streams [22,29]. Unlike CE, sample injection, separation, and collection can take place continuously because the direction of the separation is different from that of the bulk flow [22]. The continuous nature of this technique provides a highthroughput (0.1 g or more in less than an hour) [30] separation mechanism, making FFE a useful technique for preparative separations of biological molecules. Considering many similarities between GO nanosheets and biological molecules, this technique in principle should be applicable to separate GO nanosheets with equal success. However, the so-called high-throughput FFE is just available in the field of biochemistry, the throughput of this technology so far is far from the demand of large-scale preparation in the field of nanomaterials separation. In present work, we redesigned the device to make it more applicable to separate GO nanosheets as a high-efficiency preparative separation technique. By utilizing this new device, GO nanosheets are separated into different size groups with narrow size distributions. It demonstrated that, indeed, the FFE is applicable to nanoscale dispersions producing GO nanosheets with distinct sizes. The size distributions of the separated samples are monitored using atomic force microscopy (AFM). IR spectroscopy was used to study the composition of GO after separation. C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Figure 1. Schematic illustration of the FFE apparatus.
2 Materials and methods 2.1 Materials and reagents Pristine graphite was purchased from Shanghai Yifan graphite products. Other analytical grade chemical reagents were obtained from Sinopharm Chemical Reagent. GO was prepared from pristine graphite according to Hummers’ method [31].
2.2 FFE redesign Figure 1 shows the schematic illustration of FFE apparatus. As illustrated in Fig. 1, the custom apparatus of FFE is composed of high-voltage power supply (Fig. 1A), steady flow devices (Fig. 1B), peristaltic pumps (Fig. 1C), and a planar separation channel (Fig. 1D). The separation channel made of transparent organic glass is convenient for observing the separation process. The FFE electrodes (Pt electrode) are physically isolated from the separation region by IEX membranes (nitrated cellulose sheets) to prevent electrolysis bubbles from disrupting the sample streams and electric field. Carrier buffer is pumped in by the peristaltic pump through buffer inlets while GO sample is infused into the separation channel by peristaltic pump through the sample inset where the sample meets the carrier buffer and is carried all the way down the separation channel. Rubber tubes were used as the connecting channel. A high electric field is applied through the electrolyte channels positioned at both sides of the separation channel and provides the driving force for the GO samples. When the charged GO nanosheets are www.jss-journal.com
Other Techniques
J. Sep. Sci. 2015, 38, 157–163
infused into the separation channel without the electric field, they will be carried by the carrier buffer flow all the way to the end of the channel without any changes. In addition, these GO nanosheets will move in a direction parallel to the electric field with a velocity when an external electric field is applied. However, the FFE still cannot meet the requirements of large-scale preparation in the field of nanomaterials, however it has been developed as a preparative technique for biotechnology [32–35]. Thus, we did a series of experiments to optimize the FFE to make it more applicable to separate GO nanosheets on a large scale. The resolution in FFE, which largely determined the separation effect, is complicated by a number of broadening mechanisms. While redesigning the device for the separation of GO by FFE, several factors including electric field intensity, flow velocity, initial sample band, the size and volume of separation channel, etc., are needed to be considered based on a general understanding of FFE. All of these factors will affect the resolution of FFE system. There are basically two ways to achieve the purpose of large-scale preparation, increasing the injection pore diameter and flow velocity. And the injection pore diameter, flow velocity, and voltage are crucial to size separation. The change of these variables, however, will give rise to a “chain reaction.” The injection pore diameter to some extent determines the amount of sample in the continuous stream. However, the increase of the injection pore diameter will result in the increase of the initial sample bandwidth, and then decrease the resolution of separation [22]. Until now, the development of FFE is still to focus on how to narrow the bandwidth, improve the resolution, and increase the throughput. It will face the same issue when it comes to the flow velocity. The flow velocity controls the residence time of samples in the channel as well as the throughput of separation. Decreasing the flow velocity will increase the separation distance between samples [29], which was useful for improving the resolution of the separation. However, it will decrease throughput. On the contrary, the higher the flow velocity, the shorter the residence time and hence the higher the throughput of the separated samples. In this condition, the resolution of separation is sacrificed due to the shorter residence time. The voltage plays an important role in the separation process. Increasing the voltage can also increase the separation distance between samples [29]. A voltage increases, however, will also increase the amount of Joule heat generated and increase the electro-osmotic distortion. A compromise between throughput and resolution must be made by adjusting the injection pore diameter, flow velocity, and voltage. To the best of our knowledge, there has been no theoretical research on how to choose the reasonable general condition for FFE because of the intimate interaction between these factors. After a series of comparative experiments, we select 2 mm as inner diameter for injection pore, and both of the buffer flow rate and sample infusion rate are 15 mL/min. It takes as little as 1 min from the injection of the GO dispersion to collection of the fractionated samples. In addition, the size of separation channel was 170 mm × 105 mm × 2 mm, the sodium hydroxide solution C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
159
Figure 2. Change of sample bandwidth during the separation process.
(0.001 mol/L) with a pH value of 11 was selected as buffer solution, and the electric voltage in the process of electrophoresis was 1000 V. In addition, an optical image of this separation device is shown in Supporting Information Fig. S1. 2.3 Characterization The Zeta potential of the GO was measured by the Malvern Zetasizer Nano-ZS 90. The whole separation process was recorded by a digital camera. GO nanosheets were studied by AFM, and the AFM samples were prepared by depositing sufficiently diluted GO colloidal dispersions on freshly cleaved mica substrates. The AFM study was achieved with a commercial instrument (SPI3800N, Seiko Instruments, Japan), with a speed of 1 Hz. The microstructure was characterized by FTIR spectroscopy (EQUINOXSS/HYPERION2000, Bruker, Germany). Size measurement was performed using the software “Image Pro Plus.” For each sample analysis, four representative AFM images were selected.
3 Results and discussion Figure 2 shows the change of sample bandwidth during the separation process. Figure 2A is corresponding to the sample stream before the separation while Fig. 2B shows the sample stream after 20 s of applying the electric field. It was clearly evident that the bandwidth of the sample stream broadened greatly after the applied of electric field. This phenomenon should be ascribed to the migration of the sample, which was www.jss-journal.com
160
Y. Liu et al.
J. Sep. Sci. 2015, 38, 157–163
Figure 3. AFM images of GO samples and the statistical results of lateraldimensional sizes of GO nanosheets: (1) the original samples; (2) samples close to the anode; (3) samples far from the anode.
parallel to the electric field. This indicates that FFE successfully separated the GO samples into different group. A short movie presented in the Supporting Information shows the separation process. The first buzz in the movie represents the beginning of an applied electric field, while the second buzz means the end of the applied electric field. When an electric field is applied, the bandwidth broadened greatly within 10 s. And when the electric field is removed, the bandwidth can return to normal within 3 s. This phenomenon can further confirm the feasibility of FFE to separate the GO nanosheets. Figure 3 shows AFM images of GO samples and the statistical results of lateral-dimensional sizes of GO nanosheets. And the statistical results are determined based on the AFM images with the software (Image Pro Plus version 6). For
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
the original sample (1), the lateral-dimensional size varies widely in a broadly scope. After the FFE processing, however, the distribution of the lateral-dimensional size of the separated samples is more concentrated in the viewpoint of mathematic statistic. The statistical results showed that we could obtain GO sheets with a narrow variation range of the lateral-dimensional size by FFE. In addition, the average lateral-dimensional size of the samples that were closer to the anode (3) is smaller compared with the samples that were far away from the anode (2). This phenomenon indicated that the sample with smaller size has a larger deflection angle. The mechanism of this phenomenon will be discussed in the following. The GO upon exfoliation carries sufficient functional groups (e.g. epoxides, hydroxyls, carboxylic acids) on the sheet www.jss-journal.com
Other Techniques
J. Sep. Sci. 2015, 38, 157–163
161
The charge distribution on the nanosheets surface is concentrated on the edge of the GO nanosheets as carboxyl groups are mainly distributed on the edge of the GO nanosheets [36, 37] Q can be given by the following equation: Q = 2r 2 · x1 + 2r · x2 · e · q
(5)
where e is the charge density, and x1 and x2 are the percentage of charge distribution (x2 >> x1 ). We can obtain a simplified equation as all the parameters except the radius r are constant Q = K1 · r n Figure 4. Schematic drawing of the ionization of GO and ideal model for GO nanosheet.
surface (Fig. 4A). In addition, it is now widely accepted that epoxides and hydroxyl are the major functional groups that randomly distribute across the carbon layer, and at the edge are mainly carboxyls, lactones, and carbonyls [36, 37]. The ionization of the functional groups (carboxyls) will occur in the solution, and thus the GO sheets are negatively charged (Fig. 4B). The results of Zeta potential measurement also verified this view. GO sheets with various sizes possess different quantities of functional groups, which offer them with different surface charge intensities. The FFE separation method takes advantage of the different velocities between various sized GO nanosheets. However, its so complex that it is virtually impossible to analyze the movement of GO in the electric field due to its irregular shape. Thus, we use the “idealized method,” which was widely used in physical research to build an ideal model. We simplify the GO nanosheet as a platelike structure with the radius of r (Fig. 4C). According to the colloid theory, the viscous drag force f for the GO, a kind of platelike colloid, under unit electric field strength can be calculated based on the following formula [38]: f = av 2 r 2
(1)
where a is a constant, is the viscosity, r is the charged particles radius, and v is the velocity of the charges particles. The driving force (F) is given by F =E·Q
(2)
where E is the applied electric field, and Q is the charge of nanoparticles. The system will reach an equilibrium when the viscous drag force is equal to the driving force. The velocity v can be calculated based on Eqs. (1) and (2) v = (Q/E ar )
2 0.5
(3)
The electrophoretic mobility of the charged sample is given by = v/E C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
(4)
(6)
where K1 is a constant and n is a variable from 1 to 2. We can also get a simplified formula for the electrophoretic mobility based on Eqs. (3)–(6) = K 2 · r 0.5n−1
(7)
where K2 is a constant. The angle of deflection () of the sample in the electric field increases with the electric field strength and sample electrophoretic mobility, and decreases with the increasing of flow velocity of the sample [29]. The ratio of the electrophoretic velocity to the buffer velocity is tan and is given by [29] tan =
·i A1
(8)
where i is the electric current, A1 is the cross-section of the separation channel, is the specific conductance of the buffer carrier, and is the linear velocity of the buffer carrier. The electric current i can be expressed as i = −A2 F (d/d x)
|Z|c
(9)
where A2 is the area of electrode surface, F is Faraday’s constant, d/dx equals the potential gradient E, Z is the charge on the ion, and c is the concentration of the sample. The conductivity , however, is given by =F
|Z|c
(10)
Equation (9) can be simplified to i = −A2 E
(11)
By substitution, Eq. (8) becomes tan = −
A2E A1
(12)
All the factors except electrophoretic mobility are held constant during the separation process, and is the only variable. Therefore, samples with different mobilities are deflected at different angles and thus move along different directions (Fig. 2B). The separated samples are then collected www.jss-journal.com
162
J. Sep. Sci. 2015, 38, 157–163
Y. Liu et al.
2012AA030303) and Basic Research Key Program of Shanghai (no. 12JC1408600). The authors have declared no conflict of interest.
5 References [1] Novoselov, K. S., Geim, A. K., Morozov, S. V., Jiang, D., Zhang, Y., Dubonos, S. V., Grigorieva, I. V., Firsov, A. A., Science 2004, 306, 666–669. [2] Geim, A. K., Novoselov, K. S., Nat. Mater. 2007, 6, 183– 191. Figure 5. FTIR spectra of the original sample (A) and separated sample (B).
continuously at the outlet. Obviously, the velocity is inversely proportional to the radius of the GO nanosheets as we can clearly observe from Eq. (7), and the smaller GO nanosheets were expected to be collected near the anode, while the larger GO nanosheets would be collected near the cathode, which is in good agreement with the results of our study. It is well known that the functional groups (e.g. epoxides, hydroxyls, carboxylic acids) on the GO nanosheet surface affect many of its chemical and physical properties. The electric field might make some influences on the structure of the sample during the separation. To evaluate this effect, the FTIR spectra of the GO were collected on the original GO (Fig. 5A) and separated GO samples (Fig. 5B). In the spectrum of GO as shown in Fig. 5, several peaks due to OH (O–H stretching vibrations) at 3350 and 1410 cm−1 , C=O (carboxylic acid and carbonyl moieties) at 1730 cm−1 , along with a band over a range of 1000–1400 cm−1 , which arise from the C–O (1050 cm−1 ) stretching vibrations, are presented [39]. By comparing the functional groups of the samples, we conclude that the FFE technique can separate the GO nanosheets dispersions into more mono disperse samples without changing the functionality of the samples.
[3] Allen, M. J., Tung, V. C., Kaner, R. B., Chem. Rev. 2010, 110, 132–145. [4] Li, N., Liu, G., Zhen, C., Li, F., Zhang, L. L., Cheng, H. M., Adv. Funct. Mater. 2011, 21, 1717–1722. [5] Kamat, P. V., J. Phys. Chem. Lett. 2011, 2, 242–251. [6] Zhang, H., Lv, X. J., Li, Y. M., Wang, Y., Li, J. H., ACS Nano 2010, 4, 380–386. [7] Nihar, M., David, M., Xu, Z. P., Sreeprasad, T. S., Nagaraja, A., Rodriguez, A. A., Berry, V., Nat. Commun. 2012, 3, 844. [8] Ritter, K. A., Lyding, J. W., Nat. Mater. 2009, 8, 235–242. [9] Wang, X. R., Ouyang, Y. J., Li, X. L., Wang, H. L., Guo, J., Dai, H. J., Phys. Rev. Lett. 2008, 100, 206803. [10] Obradovic, B., Kotlyar, B. R., Heinz, F., Matagne, P., Rakshit, T., Giles, M. D., Stettler, M. A., Nikonov, D. E., Appl. Phys. Lett. 2006, 88, 142102. [11] Boukhvalov, D. W., Katsnelson, M. I., J. Am. Chem. Soc. 2008, 130, 10697–10701. [12] Zhang, T. Y., Zhang, D., Shen, M., Mater. Lett. 2009, 63, 2051–2054. [13] Green, A. A., Hersam, M. C., J. Phys. Chem. Lett. 2010, 1, 544–549. [14] Green, A. A., Hersam, M. C., Nano. Lett. 2009, 9, 4031– 4036. [15] Sun, X. M., Liu, Z., Welsher, K., Robinson, J. T., Goodwin, A., Zaric, S., Dai, H., Nano. Res. 2008, 1, 203–212. [16] Righetti, P. G., J. Chromatogr. A 2005, 1079, 24–40.
4 Concluding remarks
[17] Xu, X. Y., Caswell, K. K., Tucker, E., Kabisatpathy, S., Brodhacker, K. L., Scrivens, W. A., J. Chromatogr. A 2007, 1167, 35–41.
In this paper, we present a novel and simple means to separate GO nanosheets into less polydisperse size groups by utilizing the FFE technique. The capability of continuous separation of FFE gives higher yield than any other separation methods ever reported. The AFM further confirms that the separated GO nanosheets are less polydisperse. FTIR spectra show that the FFE progress does not chemically or physically modify the GO nanosheets. These results suggested that FFE has the potential to develop into an extremely powerful preparative scale separation technique in the field of nanomaterials as well as biochemistry.
[20] Zhao, J. J., Chen, G. F., Zhang, W., Li, P., Wang, L., Yue, Q. L., Wang, H. S., Dong, R. X., Yan, X. L., Liu, J. F., Anal. Chem. 2011, 83, 9100–9106.
This research was supported by National High Technology Research and Development Program of China (no.
[24] Kasicka, V., Prusik, Z., Sazelova, P., Jiracek, J., Barth, T., J. Chromatogr. A 1998, 796, 211–220.
[18] Surugau, N., Urban P. L. J., Sep. Sci. 2009, 32, 1889–1906. ¨ [19] Muller, M. B., Quirino, J. P., Nesterenko, P. N., Haddad, P. R., Gambhir, S., Li, D., Wallace, G. G., J. Chromatogr. A 2010, 1217, 7593–7597.
[21] Ho, S., Critchley, K., Lilly, G. D., Shim, B., Kotov, N. A., J. Mater. Chem. 2009, 19, 1390–1394. [22] Turgeon, R. T., Bowser, M. T., Anal. Bioanal. Chem. 2009, 394, 187–198. [23] Krivankova, L., Bocek, P., 1998, 19, 1064–1074.
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.jss-journal.com
J. Sep. Sci. 2015, 38, 157–163
Other Techniques
163
[26] Wildgruber, R., Weber, G., Wise, P., Grimm, D., Bauer, J., Proteomics 2014, 14, 629–636.
G., Nissum, M., Westphal, K., Egli, M., Grosse, J., Schonberger, J., Wildgruber, R., Infanger, M., Grimm, D., Proteomics 2010, 10, 904–913.
[27] Dong, Y. C., Shao, J., Yin, X. Y., Fan, L. Y., Cao, C. X., J. Sep. Sci. 2011, 34, 1683–1691.
[34] Weber, G., Islinger, M., Weber, P., Eckerskorn, C., Volkl, A., Electrophoresis 2004, 25, 1735–1747.
[28] Wang, P. L., Zhang, L. H., Shan, Y. C., Cong, Y. Z., Liang, Y., Han, B., Liang, Z., Zhang, Y. K., J. Sep. Sci., 2010, 33, 2039–2044.
[35] Zischka, H., Weber, G., Weber P. J. A., Posch, A., Braun, R. J., Buhringer, D., Schneider, U., Nissum, M., Meitinger, T., Ueffing, M., Eckerskorn, C., Proteomics 2003, 3, 906– 916.
[25] Kasicka, V., Electrophoresis 2009, 30, S40–S52.
[29] Roman, M. C., Brown, P. R., Anal. Chem. 1994, 66, 86–94. [30] Peterson, R. R., Cliffel, D. E., Anal. Chem. 2005, 77, 4348– 4353. [31] Hummers, W. S., Offeman, R. E., J. Am. Chem. Soc. 1958, 80, 1339–1339. [32] Ma, X., Wildgruber, R., Bauer, J., Weber, G., Infanger, M., Grosse, J., Grimm, D., Electrophoresis 2012, 33, 1349– 1355. [33] Pietsch, J., Kussian, R., Sickmann, A., Bauer, J., Weber,
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
[36] Lerf, A., He, H., Forster, M., Klinowski, J., J. Phys. Chem. B 1998, 102, 4477–4482. [37] Mao, S., Pu, H. H., Chen, J. H., RSC Adv. 2012, 2, 2643– 2662. [38] Yuan, Z. W., Chu, F. L., Wang, S. B., J. Vib. Eng. 2004, 17, 6–8. [39] Yu, D., Yang, Y., Durstock, M., Baek, J.-B., Dai, L., ACS Nano 2010, 4, 5633–5640.
www.jss-journal.com
J. Sep. Sci. 2015, 38, 157–163
Yong Liu Dong Zhang Shiwu Pang Yanyun Liu Yu Shang Key Laboratory of Advanced Civil Engineering Materials, School of Materials Science and Engineering, Tongji University, Shanghai, P. R. China Received September 11, 2014 Revised October 17, 2014 Accepted October 20, 2014
Research Article
Size separation of graphene oxide using preparative free-flow electrophoresis Graphene oxide nanosheets often bear a wide size distribution. However, it is critical to have nanosheets with narrow size distribution for their unique size-dependent physiochemical properties, and nanosheets with a narrow size distribution are the cornerstones for application. Therefore, efficient separation methods of graphene nanosheets have been given considerable attention in many scientific areas recently. Free-flow electrophoresis is extensively used in the separation and purification of biological molecules with continuous flow separation. The charged graphene oxide nanosheets to some extent are very close in size to biological molecules and share similarity in motion behavior in an electric field. Thus, in the present work, we present a new and simple means to separate graphene oxide nanosheets into more mono-dispersed size groups by using the free-flow electrophoresis technique. By optimizing the separation conditions, we were able to obtain graphene oxide sheets with narrow size distribution. The separated samples were characterized by atomic force microscopy, and the size measurements were made by using the software “Image Pro Plus.” In addition, a brief discussion is also given into the theoretic background of the separation of graphene oxide according to the size by the technique of preparative free-flow electrophoresis.
Keywords: Free-flow electrophoresis / Graphene oxide DOI 10.1002/jssc.201401000
Additional supporting information may be found in the online version of this article at the publisher’s web-site
1 Introduction Graphene, a single layer of carbon atoms densely packed into a benzene-ring structure [1], has received considerable attention due to its unique electronic properties, high transparency, flexible structure, large specific surface area, and good thermal conduction [2, 3] since Geim et al. [1] made a breakthrough in the preparation of graphene by mechanical exfoliation. The peculiar chemical and physical properties of graphene are of great interest for potential applications in various fields, such as composite materials, batteries, sensors, solar cells and photo catalysts, and many others [4–6]. It is well known that the shape and size of graphene nanostructure dictate the electrical, optical, magnetic, and chemical properties because of the edge states and quantum Correspondence: Professor Dong Zhang, Key Laboratory of Advanced Civil Engineering Materials, School of Materials Science and Engineering, Tongji University, 4800 CaoAn Road, Shanghai 201804, P. R. China E-mail: [email protected] Fax: +86-21-69582144
Abbreviations: AFM, atomic force microscopy; FFE, free-flow electrophoresis; GO, graphene oxide
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
confinement [7–10]. Therefore, the key to controlling the properties mostly lies in the control of the size and shape of graphene. Unfortunately, current preparative processes of graphene have generally not generated a uniform size of nanosheets. In addition, graphene is poorly soluble in water or polar organic solvents, which makes it more difficult to further process to control the size and shape of the graphene nanosheets. Unlike hydrophobic graphene, graphene oxide (GO) is hydrophilic due to the oxygencontaining functional groups on the sheet surface, which gives the GO sheets good solubility in solvents and provides fertile opportunities for the processability of this nanomaterial [11]. Furthermore, GO carries sufficient functional groups (e.g. epoxides, hydroxyls, carboxylic acids) on the sheet surface. Ionization of the functional groups will occur in solution, and thus the GO sheets are negatively charged. In addition, GO has attracted intense interest for its use as a precursor material for the mass production of graphene-based materials, which hold great potential in various applications. Considering their unique size and/or shape-dependent properties, it is highly important to obtain GO nanosheets with a narrow size and shape dispersion. Many attempts have been made toward solving the above-mentioned issues such as magnetic stirring and
www.jss-journal.com
158
J. Sep. Sci. 2015, 38, 157–163
Y. Liu et al.
centrifugation [12], isopycnic density gradient ultracentrifugation [13, 14], and the density gradient ultracentrifugation rate [15]. However, it is not favorable to use centrifugation for separating GO on a large scale only if super-magnum centrifuge is given. It would be laborious and time-consuming if large amounts of separated samples are needed as it is done on a batch-by-batch basis, and lacks continuous separation capacity. Electrophoresis has been extensively used in the separation and purification of biological molecules with remarkable resolution [16]. Surprisingly, this technology has gained little exposure until recently in the separation of nanomaterials [17,18]. CE, a liquid-phase separation technique, has been reported for the separation of GO [19, 20]. In this technique, capillary as separation channel (micron scale), and high direct current (DC) electric voltage (about 30 kV) as the main driving force together achieve the analysis; CE lacks continuous separation capacity as separation is done on a batch-by-batch basis [21]. The CE method is done on a miniscule scale. Therefore, finding a fast, effective, and continuous method of GO separation remains a practical problem of critical importance. Free-flow electrophoresis (FFE), another little known electrophoresis technique, has evolved into one of the most promising techniques for preparative-scale separation in the field of biochemical technology [21–28]. FFE is a pure-liquid electrophoresis for both analytical and preparative separation in biochemical technology with the advantages of controllability, gentle conditions, high recovery, and complete preservation of activity. In FFE, pressure is used to drive a sample stream through a planar separation channel. An electric field is applied perpendicularly to the direction of carrier buffer flow while a sample solution is injected continuously into the carrier buffer as a narrow band to deflect sample into distinct streams [22,29]. Unlike CE, sample injection, separation, and collection can take place continuously because the direction of the separation is different from that of the bulk flow [22]. The continuous nature of this technique provides a highthroughput (0.1 g or more in less than an hour) [30] separation mechanism, making FFE a useful technique for preparative separations of biological molecules. Considering many similarities between GO nanosheets and biological molecules, this technique in principle should be applicable to separate GO nanosheets with equal success. However, the so-called high-throughput FFE is just available in the field of biochemistry, the throughput of this technology so far is far from the demand of large-scale preparation in the field of nanomaterials separation. In present work, we redesigned the device to make it more applicable to separate GO nanosheets as a high-efficiency preparative separation technique. By utilizing this new device, GO nanosheets are separated into different size groups with narrow size distributions. It demonstrated that, indeed, the FFE is applicable to nanoscale dispersions producing GO nanosheets with distinct sizes. The size distributions of the separated samples are monitored using atomic force microscopy (AFM). IR spectroscopy was used to study the composition of GO after separation. C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Figure 1. Schematic illustration of the FFE apparatus.
2 Materials and methods 2.1 Materials and reagents Pristine graphite was purchased from Shanghai Yifan graphite products. Other analytical grade chemical reagents were obtained from Sinopharm Chemical Reagent. GO was prepared from pristine graphite according to Hummers’ method [31].
2.2 FFE redesign Figure 1 shows the schematic illustration of FFE apparatus. As illustrated in Fig. 1, the custom apparatus of FFE is composed of high-voltage power supply (Fig. 1A), steady flow devices (Fig. 1B), peristaltic pumps (Fig. 1C), and a planar separation channel (Fig. 1D). The separation channel made of transparent organic glass is convenient for observing the separation process. The FFE electrodes (Pt electrode) are physically isolated from the separation region by IEX membranes (nitrated cellulose sheets) to prevent electrolysis bubbles from disrupting the sample streams and electric field. Carrier buffer is pumped in by the peristaltic pump through buffer inlets while GO sample is infused into the separation channel by peristaltic pump through the sample inset where the sample meets the carrier buffer and is carried all the way down the separation channel. Rubber tubes were used as the connecting channel. A high electric field is applied through the electrolyte channels positioned at both sides of the separation channel and provides the driving force for the GO samples. When the charged GO nanosheets are www.jss-journal.com
Other Techniques
J. Sep. Sci. 2015, 38, 157–163
infused into the separation channel without the electric field, they will be carried by the carrier buffer flow all the way to the end of the channel without any changes. In addition, these GO nanosheets will move in a direction parallel to the electric field with a velocity when an external electric field is applied. However, the FFE still cannot meet the requirements of large-scale preparation in the field of nanomaterials, however it has been developed as a preparative technique for biotechnology [32–35]. Thus, we did a series of experiments to optimize the FFE to make it more applicable to separate GO nanosheets on a large scale. The resolution in FFE, which largely determined the separation effect, is complicated by a number of broadening mechanisms. While redesigning the device for the separation of GO by FFE, several factors including electric field intensity, flow velocity, initial sample band, the size and volume of separation channel, etc., are needed to be considered based on a general understanding of FFE. All of these factors will affect the resolution of FFE system. There are basically two ways to achieve the purpose of large-scale preparation, increasing the injection pore diameter and flow velocity. And the injection pore diameter, flow velocity, and voltage are crucial to size separation. The change of these variables, however, will give rise to a “chain reaction.” The injection pore diameter to some extent determines the amount of sample in the continuous stream. However, the increase of the injection pore diameter will result in the increase of the initial sample bandwidth, and then decrease the resolution of separation [22]. Until now, the development of FFE is still to focus on how to narrow the bandwidth, improve the resolution, and increase the throughput. It will face the same issue when it comes to the flow velocity. The flow velocity controls the residence time of samples in the channel as well as the throughput of separation. Decreasing the flow velocity will increase the separation distance between samples [29], which was useful for improving the resolution of the separation. However, it will decrease throughput. On the contrary, the higher the flow velocity, the shorter the residence time and hence the higher the throughput of the separated samples. In this condition, the resolution of separation is sacrificed due to the shorter residence time. The voltage plays an important role in the separation process. Increasing the voltage can also increase the separation distance between samples [29]. A voltage increases, however, will also increase the amount of Joule heat generated and increase the electro-osmotic distortion. A compromise between throughput and resolution must be made by adjusting the injection pore diameter, flow velocity, and voltage. To the best of our knowledge, there has been no theoretical research on how to choose the reasonable general condition for FFE because of the intimate interaction between these factors. After a series of comparative experiments, we select 2 mm as inner diameter for injection pore, and both of the buffer flow rate and sample infusion rate are 15 mL/min. It takes as little as 1 min from the injection of the GO dispersion to collection of the fractionated samples. In addition, the size of separation channel was 170 mm × 105 mm × 2 mm, the sodium hydroxide solution C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
159
Figure 2. Change of sample bandwidth during the separation process.
(0.001 mol/L) with a pH value of 11 was selected as buffer solution, and the electric voltage in the process of electrophoresis was 1000 V. In addition, an optical image of this separation device is shown in Supporting Information Fig. S1. 2.3 Characterization The Zeta potential of the GO was measured by the Malvern Zetasizer Nano-ZS 90. The whole separation process was recorded by a digital camera. GO nanosheets were studied by AFM, and the AFM samples were prepared by depositing sufficiently diluted GO colloidal dispersions on freshly cleaved mica substrates. The AFM study was achieved with a commercial instrument (SPI3800N, Seiko Instruments, Japan), with a speed of 1 Hz. The microstructure was characterized by FTIR spectroscopy (EQUINOXSS/HYPERION2000, Bruker, Germany). Size measurement was performed using the software “Image Pro Plus.” For each sample analysis, four representative AFM images were selected.
3 Results and discussion Figure 2 shows the change of sample bandwidth during the separation process. Figure 2A is corresponding to the sample stream before the separation while Fig. 2B shows the sample stream after 20 s of applying the electric field. It was clearly evident that the bandwidth of the sample stream broadened greatly after the applied of electric field. This phenomenon should be ascribed to the migration of the sample, which was www.jss-journal.com
160
Y. Liu et al.
J. Sep. Sci. 2015, 38, 157–163
Figure 3. AFM images of GO samples and the statistical results of lateraldimensional sizes of GO nanosheets: (1) the original samples; (2) samples close to the anode; (3) samples far from the anode.
parallel to the electric field. This indicates that FFE successfully separated the GO samples into different group. A short movie presented in the Supporting Information shows the separation process. The first buzz in the movie represents the beginning of an applied electric field, while the second buzz means the end of the applied electric field. When an electric field is applied, the bandwidth broadened greatly within 10 s. And when the electric field is removed, the bandwidth can return to normal within 3 s. This phenomenon can further confirm the feasibility of FFE to separate the GO nanosheets. Figure 3 shows AFM images of GO samples and the statistical results of lateral-dimensional sizes of GO nanosheets. And the statistical results are determined based on the AFM images with the software (Image Pro Plus version 6). For
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
the original sample (1), the lateral-dimensional size varies widely in a broadly scope. After the FFE processing, however, the distribution of the lateral-dimensional size of the separated samples is more concentrated in the viewpoint of mathematic statistic. The statistical results showed that we could obtain GO sheets with a narrow variation range of the lateral-dimensional size by FFE. In addition, the average lateral-dimensional size of the samples that were closer to the anode (3) is smaller compared with the samples that were far away from the anode (2). This phenomenon indicated that the sample with smaller size has a larger deflection angle. The mechanism of this phenomenon will be discussed in the following. The GO upon exfoliation carries sufficient functional groups (e.g. epoxides, hydroxyls, carboxylic acids) on the sheet www.jss-journal.com
Other Techniques
J. Sep. Sci. 2015, 38, 157–163
161
The charge distribution on the nanosheets surface is concentrated on the edge of the GO nanosheets as carboxyl groups are mainly distributed on the edge of the GO nanosheets [36, 37] Q can be given by the following equation: Q = 2r 2 · x1 + 2r · x2 · e · q
(5)
where e is the charge density, and x1 and x2 are the percentage of charge distribution (x2 >> x1 ). We can obtain a simplified equation as all the parameters except the radius r are constant Q = K1 · r n Figure 4. Schematic drawing of the ionization of GO and ideal model for GO nanosheet.
surface (Fig. 4A). In addition, it is now widely accepted that epoxides and hydroxyl are the major functional groups that randomly distribute across the carbon layer, and at the edge are mainly carboxyls, lactones, and carbonyls [36, 37]. The ionization of the functional groups (carboxyls) will occur in the solution, and thus the GO sheets are negatively charged (Fig. 4B). The results of Zeta potential measurement also verified this view. GO sheets with various sizes possess different quantities of functional groups, which offer them with different surface charge intensities. The FFE separation method takes advantage of the different velocities between various sized GO nanosheets. However, its so complex that it is virtually impossible to analyze the movement of GO in the electric field due to its irregular shape. Thus, we use the “idealized method,” which was widely used in physical research to build an ideal model. We simplify the GO nanosheet as a platelike structure with the radius of r (Fig. 4C). According to the colloid theory, the viscous drag force f for the GO, a kind of platelike colloid, under unit electric field strength can be calculated based on the following formula [38]: f = av 2 r 2
(1)
where a is a constant, is the viscosity, r is the charged particles radius, and v is the velocity of the charges particles. The driving force (F) is given by F =E·Q
(2)
where E is the applied electric field, and Q is the charge of nanoparticles. The system will reach an equilibrium when the viscous drag force is equal to the driving force. The velocity v can be calculated based on Eqs. (1) and (2) v = (Q/E ar )
2 0.5
(3)
The electrophoretic mobility of the charged sample is given by = v/E C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
(4)
(6)
where K1 is a constant and n is a variable from 1 to 2. We can also get a simplified formula for the electrophoretic mobility based on Eqs. (3)–(6) = K 2 · r 0.5n−1
(7)
where K2 is a constant. The angle of deflection () of the sample in the electric field increases with the electric field strength and sample electrophoretic mobility, and decreases with the increasing of flow velocity of the sample [29]. The ratio of the electrophoretic velocity to the buffer velocity is tan and is given by [29] tan =
·i A1
(8)
where i is the electric current, A1 is the cross-section of the separation channel, is the specific conductance of the buffer carrier, and is the linear velocity of the buffer carrier. The electric current i can be expressed as i = −A2 F (d/d x)
|Z|c
(9)
where A2 is the area of electrode surface, F is Faraday’s constant, d/dx equals the potential gradient E, Z is the charge on the ion, and c is the concentration of the sample. The conductivity , however, is given by =F
|Z|c
(10)
Equation (9) can be simplified to i = −A2 E
(11)
By substitution, Eq. (8) becomes tan = −
A2E A1
(12)
All the factors except electrophoretic mobility are held constant during the separation process, and is the only variable. Therefore, samples with different mobilities are deflected at different angles and thus move along different directions (Fig. 2B). The separated samples are then collected www.jss-journal.com
162
J. Sep. Sci. 2015, 38, 157–163
Y. Liu et al.
2012AA030303) and Basic Research Key Program of Shanghai (no. 12JC1408600). The authors have declared no conflict of interest.
5 References [1] Novoselov, K. S., Geim, A. K., Morozov, S. V., Jiang, D., Zhang, Y., Dubonos, S. V., Grigorieva, I. V., Firsov, A. A., Science 2004, 306, 666–669. [2] Geim, A. K., Novoselov, K. S., Nat. Mater. 2007, 6, 183– 191. Figure 5. FTIR spectra of the original sample (A) and separated sample (B).
continuously at the outlet. Obviously, the velocity is inversely proportional to the radius of the GO nanosheets as we can clearly observe from Eq. (7), and the smaller GO nanosheets were expected to be collected near the anode, while the larger GO nanosheets would be collected near the cathode, which is in good agreement with the results of our study. It is well known that the functional groups (e.g. epoxides, hydroxyls, carboxylic acids) on the GO nanosheet surface affect many of its chemical and physical properties. The electric field might make some influences on the structure of the sample during the separation. To evaluate this effect, the FTIR spectra of the GO were collected on the original GO (Fig. 5A) and separated GO samples (Fig. 5B). In the spectrum of GO as shown in Fig. 5, several peaks due to OH (O–H stretching vibrations) at 3350 and 1410 cm−1 , C=O (carboxylic acid and carbonyl moieties) at 1730 cm−1 , along with a band over a range of 1000–1400 cm−1 , which arise from the C–O (1050 cm−1 ) stretching vibrations, are presented [39]. By comparing the functional groups of the samples, we conclude that the FFE technique can separate the GO nanosheets dispersions into more mono disperse samples without changing the functionality of the samples.
[3] Allen, M. J., Tung, V. C., Kaner, R. B., Chem. Rev. 2010, 110, 132–145. [4] Li, N., Liu, G., Zhen, C., Li, F., Zhang, L. L., Cheng, H. M., Adv. Funct. Mater. 2011, 21, 1717–1722. [5] Kamat, P. V., J. Phys. Chem. Lett. 2011, 2, 242–251. [6] Zhang, H., Lv, X. J., Li, Y. M., Wang, Y., Li, J. H., ACS Nano 2010, 4, 380–386. [7] Nihar, M., David, M., Xu, Z. P., Sreeprasad, T. S., Nagaraja, A., Rodriguez, A. A., Berry, V., Nat. Commun. 2012, 3, 844. [8] Ritter, K. A., Lyding, J. W., Nat. Mater. 2009, 8, 235–242. [9] Wang, X. R., Ouyang, Y. J., Li, X. L., Wang, H. L., Guo, J., Dai, H. J., Phys. Rev. Lett. 2008, 100, 206803. [10] Obradovic, B., Kotlyar, B. R., Heinz, F., Matagne, P., Rakshit, T., Giles, M. D., Stettler, M. A., Nikonov, D. E., Appl. Phys. Lett. 2006, 88, 142102. [11] Boukhvalov, D. W., Katsnelson, M. I., J. Am. Chem. Soc. 2008, 130, 10697–10701. [12] Zhang, T. Y., Zhang, D., Shen, M., Mater. Lett. 2009, 63, 2051–2054. [13] Green, A. A., Hersam, M. C., J. Phys. Chem. Lett. 2010, 1, 544–549. [14] Green, A. A., Hersam, M. C., Nano. Lett. 2009, 9, 4031– 4036. [15] Sun, X. M., Liu, Z., Welsher, K., Robinson, J. T., Goodwin, A., Zaric, S., Dai, H., Nano. Res. 2008, 1, 203–212. [16] Righetti, P. G., J. Chromatogr. A 2005, 1079, 24–40.
4 Concluding remarks
[17] Xu, X. Y., Caswell, K. K., Tucker, E., Kabisatpathy, S., Brodhacker, K. L., Scrivens, W. A., J. Chromatogr. A 2007, 1167, 35–41.
In this paper, we present a novel and simple means to separate GO nanosheets into less polydisperse size groups by utilizing the FFE technique. The capability of continuous separation of FFE gives higher yield than any other separation methods ever reported. The AFM further confirms that the separated GO nanosheets are less polydisperse. FTIR spectra show that the FFE progress does not chemically or physically modify the GO nanosheets. These results suggested that FFE has the potential to develop into an extremely powerful preparative scale separation technique in the field of nanomaterials as well as biochemistry.
[20] Zhao, J. J., Chen, G. F., Zhang, W., Li, P., Wang, L., Yue, Q. L., Wang, H. S., Dong, R. X., Yan, X. L., Liu, J. F., Anal. Chem. 2011, 83, 9100–9106.
This research was supported by National High Technology Research and Development Program of China (no.
[24] Kasicka, V., Prusik, Z., Sazelova, P., Jiracek, J., Barth, T., J. Chromatogr. A 1998, 796, 211–220.
[18] Surugau, N., Urban P. L. J., Sep. Sci. 2009, 32, 1889–1906. ¨ [19] Muller, M. B., Quirino, J. P., Nesterenko, P. N., Haddad, P. R., Gambhir, S., Li, D., Wallace, G. G., J. Chromatogr. A 2010, 1217, 7593–7597.
[21] Ho, S., Critchley, K., Lilly, G. D., Shim, B., Kotov, N. A., J. Mater. Chem. 2009, 19, 1390–1394. [22] Turgeon, R. T., Bowser, M. T., Anal. Bioanal. Chem. 2009, 394, 187–198. [23] Krivankova, L., Bocek, P., 1998, 19, 1064–1074.
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.jss-journal.com
J. Sep. Sci. 2015, 38, 157–163
Other Techniques
163
[26] Wildgruber, R., Weber, G., Wise, P., Grimm, D., Bauer, J., Proteomics 2014, 14, 629–636.
G., Nissum, M., Westphal, K., Egli, M., Grosse, J., Schonberger, J., Wildgruber, R., Infanger, M., Grimm, D., Proteomics 2010, 10, 904–913.
[27] Dong, Y. C., Shao, J., Yin, X. Y., Fan, L. Y., Cao, C. X., J. Sep. Sci. 2011, 34, 1683–1691.
[34] Weber, G., Islinger, M., Weber, P., Eckerskorn, C., Volkl, A., Electrophoresis 2004, 25, 1735–1747.
[28] Wang, P. L., Zhang, L. H., Shan, Y. C., Cong, Y. Z., Liang, Y., Han, B., Liang, Z., Zhang, Y. K., J. Sep. Sci., 2010, 33, 2039–2044.
[35] Zischka, H., Weber, G., Weber P. J. A., Posch, A., Braun, R. J., Buhringer, D., Schneider, U., Nissum, M., Meitinger, T., Ueffing, M., Eckerskorn, C., Proteomics 2003, 3, 906– 916.
[25] Kasicka, V., Electrophoresis 2009, 30, S40–S52.
[29] Roman, M. C., Brown, P. R., Anal. Chem. 1994, 66, 86–94. [30] Peterson, R. R., Cliffel, D. E., Anal. Chem. 2005, 77, 4348– 4353. [31] Hummers, W. S., Offeman, R. E., J. Am. Chem. Soc. 1958, 80, 1339–1339. [32] Ma, X., Wildgruber, R., Bauer, J., Weber, G., Infanger, M., Grosse, J., Grimm, D., Electrophoresis 2012, 33, 1349– 1355. [33] Pietsch, J., Kussian, R., Sickmann, A., Bauer, J., Weber,
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
[36] Lerf, A., He, H., Forster, M., Klinowski, J., J. Phys. Chem. B 1998, 102, 4477–4482. [37] Mao, S., Pu, H. H., Chen, J. H., RSC Adv. 2012, 2, 2643– 2662. [38] Yuan, Z. W., Chu, F. L., Wang, S. B., J. Vib. Eng. 2004, 17, 6–8. [39] Yu, D., Yang, Y., Durstock, M., Baek, J.-B., Dai, L., ACS Nano 2010, 4, 5633–5640.
www.jss-journal.com
